Is the `visitor_home_cbg` affected by worker visits?

Hi everyone, is the visitor_home_cbg affected by worker visits? If so, can you suggest best practices on how to filter out workers from this variable? In the places schema, using bucketed_dwell_time is suggested to filter workers. But which dwell_time do you suggest we use and how would one attribute workers to visitor_home_cbg?

@Alex_Zentefis_Yale_University the visitor_home_cbg column uses an algorithm that looks through up to 6 weeks of user data to determine the home cbg of the individual. I don’t believe there is a need for you to filter out the records using the medium dwell time.

Let me know if you have any more questions.

Hey @Alex_Zentefis_Yale_University , hope you are doing well! To add to Pranav’s answer, I would suggest checking out this section of the Places Manual. In short—yes, workers are included in all visit counts. As you’ve said, bucketed_dwell_time can be used to try filtering out worker. It may depend on the POI type (which could impact how long a worker’s shift would last), but generally you could use the “>240” bucket, which corresponds to visits that last more than four hours.

For example, suppose you are looking at banks. The vast majority of non-worker visits to banks will last far less than 240 minutes (4 hours).

I would also strongly suggest you look at the visitor_daytime_cbgs column. Check it out here in the Places Manual. It is like visitor_home_cbgs, but aggregates visitors by their typical daytime location instead of nighttime, which may be helpful in determining work behavior.

Finally, Neighborhood Patterns has a workday_device_home_areas column that may be useful.

Ok! Thank you @Pranav_Thaenraj_SafeGraph and @Ryan_Kruse_MN_State!

Hi @Ryan_Kruse_MN_State, I don’t see a description of workday_device_home_areas in the Neighborhoods documentation. Was this variable split up into work_hours_device_home_areas and work_behavior_device_home_areas ?

@Alex_Zentefis_Yale_University Sorry—yes, those are the two variables you should be looking at :slightly_smiling_face: